High-precision surface-profiling devices of the past have been primarily of the contact type. A spherical ruby stylus (of radius typically 1 mm) is used to probe he surface under test. In some systems, the deflection of the probe is measured by the closure of a mechanical microswitch; in other systems, the deflection is measured capacitively. Although such probes can be exceedingly accurate on suitable materials, the range of suitable materials is limited. Inaccurate readings or damage may occur if the sample is unsuitable. Examples of unsuitable samples include: soft, liquid, or sticky materials, very thin samples that would deform under stylus pressure, samples with concavities of radius smaller than the tip radius, or samples that are hazardous to touch. Additionally, in order to minimize probe and sample wear, the speed of measurement must be made low.
As an alternative to the electro-mechanical controllers and ruby-tipped contact probes, several types of measurement systems have been developed using optical triangulation. U.S. Pat. No. 4,733,969 discloses a laser probe for determining distance from the probe to the test object by projecting laser light onto a target surface through a source lens. The light is reflected from the target surface through a receiving lens and directed onto a pair of detectors. Light falling on the first detector indicates that the sensor (or probe) is within range for making a measurement. Light falling equally on, or in a predetermined proportion on, the two detectors indicates that a trigger point is reached and a coordinate measurement should be taken. Light falling off the second detector indicates that the sensor is no longer in range for measurements. The probe is able to focus the laser beam to approximately 0.001 inches. Alternatively, a cylindrical lens may be incorporated to project a stripe pattern onto the target object. The laser probe may be produced in a compact configuration using a reflecting mirror and additional focusing lenses. The advantages of this laser probe include the ability to integrate with existing coordinate measuring machines, the production of a very small spot size for measurement of very small and detailed objects, and a probe response speed that is fast and accurate. Unfortunately, this device only provides a binary comparison of distance against a pre-determined reference, whereas a numerical range reading would be more generally useful.
U.S. Pat. No. 4,891,772 and 4,872,747 describe a point and line range sensor that produces a numerical range reading based on optical triangulation. Once again, laser light is directed onto a target surface. The light reflected from the target surface is passed through a collimating lens to create a spot of light on a suitable multi-element detector, and the position of the spot of light is analyzed.
A key feature of the apparatus disclosed in U.S. Pat. No. 4,891,772 and 4,872,747 is the use of prisms to produce anamorphic magnification. This technique makes the instrument much more compact, while allowing the light to remain concentrated, which improves the performance of the instrument. After the returning light has passed through the collimating lens, it is directed toward the roughly-perpendicular face of a prism. The light exits the prism at a steep angle providing large angular magnification. A focusing lens directs the exiting collimated light onto the surface of a detector and a range measurement is determined. Alternatively, and preferably, two prisms are used, each providing approximately equal magnification. This scheme provides better performance than a single prism, since the dispersion and distortion from the first prism can be largely canceled by opposite dispersion and distortion from the second prism. The second prism may be oriented to provide total internal reflection, which can further reduce package size. The advantages of this system include the avoidance of contact with the target object, a small package size and the ability to maintain post-magnification light levels at substantially higher levels than non-anamorphic systems.
At least one limitation of the apparatus of U.S. Pat. Nos. 4,891,772 and 4,872,747 is that the anamorphic magnification results in considerable field-dependent astigmatism. This causes the spot on the detector to widen toward the ends of the working range of the sensor, which increases the uncertainty of the spot location. The astigmatism arises from fundamental characteristics of image formation from tilted planes, and cannot be reduced by any simple optimization of surface radii and tilts.
U.S. Pat. No. 5,633,721 describes a surface-position detection apparatus that uses optical triangulation to measure the location of a pattern of light directed onto a target surface. The return light from the pattern is collected and analyzed by a receiving optical system, consisting of a focusing lens, a prism or grating, a relay lens, and a spatially-sensitive detector. The focusing lens forms an image of the pattern on the surface of the prism or grating, and this image is re-focused onto the detector by the relay lens. The target surface, the focusing lens, and the front prism surface are tilted so as to satisfy the Scheimpflug condition. The prism serves to refract the image of the pattern and to reduce the obliquity of the final image.
Limitations on the technology disclosed in U.S. Pat. No. 5,633,721 relate to the prism. The front surface of the prism lies in an image plane and as such, any dust or surface defects on the prism appear in sharp focus on the detector, and can easily cause objectionable image degradation. Additionally, because the image is formed on the front surface of the prism an additional optical device (a relay lens) must be included, which adds to the bulk, weight and complexity in assembling the sensor system.
Moreover, each of the above patents faces the problem of limited range resolution. A fundamental limit on range resolution is imposed by the presence of speckle. Speckle is a well-known interference phenomenon observed with coherent sources, where irregularities in a surface placed in the optical path give rise to a characteristic granular spatial variation in the light intensity. Since this granularity persists despite focusing of the light, it places a fundamental limit on the resolution of laser triangulation sensors. The phenomenon of speckle, as it relates to optical triangulation, was analyzed in a paper entitled "Laser Triangulation; Fundamental Uncertainty in Distance Measurement by Rainer G. Dorsch, Gerd Hausler, and Jurgen Herrmann (Applied Optics, Vol. 33 (1994) pp. 1306-1314). The conclusion reached in the analysis was that, in order to reduce or eliminate the phenomenon of speckle, the object-space numerical aperture of the receiver lens must be maximized for optimal performance. The above-described surface measurement system patents did not address the maximization of numerical aperture. Due to neglect of this important factor, earlier sensors had limited range-to-resolution ratios, typically in the range of only 400:1.
Further, each of the above-described patents is only able to take and process a single surface measurement at a time. In other words, each of the above-described optical systems is designed so that the detector may detect only one spot of light, that is reflected from the target surface at a time.
In view of the above, there is a need for an optical sensor system that can produce precise measurements by reducing speckle through numerical aperture adjustment, that has improved range-to-resolution and range-to-accuracy ratios and that can determine multiple spots of light on the detector, which would enable it to make thickness measurements of transparent objects.